Protein hypergraph construction method, construction device and equipment

Abstract

The present application discloses a method for constructing a protein hypergraph, a device, a medium, a device and a program product for constructing a protein hypergraph. The method can be applied to the field of protein representation learning. The method comprises: constructing a plurality of structural hyperedges according to the spatial relationship between all amino acids contained in the protein to obtain a first hyperedge set of the protein, wherein the structural hyperedges contain a plurality of amino acids; obtaining the amino acid features of each amino acid, and constructing a vertex feature vector of each amino acid according to the amino acid features to obtain a vertex feature set of the protein; and determining the protein hypergraph according to the first hyperedge set and the vertex feature set. The method of the present application can enable the generated protein hypergraph to better represent the high-order information in the protein.

Description

Technical Field

[0001] The present application relates to the field of computer technology, and in particular to a method for constructing a protein hypergraph, a device, medium, equipment and program product for constructing a protein hypergraph. Background Art

[0002] Proteins are biological macromolecules with 3D spatial structures composed of folded and coiled amino acids, and their functions also depend on spatial structures. Among them, protein representation learning is an important branch in the field of protein research. Existing protein representation learning methods are based on sequences and graphs. For example, graph-based protein representation learning uses a certain model or method to convert entities and relationships in the graph into a unified vector representation of a certain dimension. However, using sequences and graphs to model proteins essentially ignores the interaction relationship between multiple amino acids in space, and thus cannot well represent the high-order relationship of proteins. Summary of the invention

[0003] The embodiments of the present application provide a method for constructing a protein hypergraph, a device, medium, equipment and program product for constructing a protein hypergraph, so that the generated protein hypergraph can better represent the high-order information in the protein.

[0004] In one aspect, a method for constructing a protein hypergraph is provided, comprising:

[0005] Constructing a plurality of structural hyperedges according to the spatial relationship between all amino acids contained in the protein to obtain a first hyperedge set of the protein, wherein the structural hyperedges contain a plurality of amino acids;

[0006] Acquire the amino acid feature of each amino acid, and construct the vertex feature vector of each amino acid according to the amino acid feature to obtain the vertex feature set of the protein;

[0007] A protein hypergraph is determined according to the first hyperedge set and the vertex feature set.

[0008] In another aspect, a device for constructing a protein hypergraph is provided, comprising:

[0009] A first construction unit is used to construct a plurality of structural hyperedges according to a plurality of amino acids of a protein to obtain a first hyperedge set of the protein, wherein the structural hyperedges include a plurality of vertex amino acids;

[0010] The second construction unit is used to obtain the amino acid feature of each vertex amino acid, and construct a vertex feature vector of each vertex amino acid according to the amino acid feature to obtain a vertex feature set of the protein;

[0011] A determining unit is used to determine a protein hypergraph according to the first hyperedge set and the vertex feature set.

[0012] On the other hand, a computer-readable storage medium is provided, wherein the computer-readable storage medium stores a computer program, and the computer program is suitable for being loaded by a processor to execute the steps in the method of any of the above embodiments.

[0013] On the other hand, a computer device is provided. The computer device includes a processor and a memory. A computer program is stored in the memory. The processor executes the steps in the method of any of the above embodiments by calling the computer program stored in the memory.

[0014] On the other hand, a computer program product is provided, comprising computer instructions, which implement the steps in the method of any of the above embodiments when the computer instructions are executed by a processor.

[0015] The embodiment of the present application constructs multiple structural hyperedges according to the spatial relationship between all the amino acids contained in the protein to obtain the first hyperedge set of the protein, and the structural hyperedges contain multiple amino acids; obtains the amino acid feature of each amino acid, and constructs the vertex feature vector of each amino acid according to the amino acid feature to obtain the vertex feature set of the protein; determines the protein hypergraph according to the first hyperedge set and the vertex feature set. The protein hypergraph construction method provided in the embodiment of the present application constructs multiple amino acid interaction sequence structural hyperedges and spatial structural hyperedges in space, and also incorporates the feature information of the vertex amino acids, so that the constructed hypergraph can reflect the spatial relationship and amino acid features between amino acids, and can model the high-order information of the protein to a certain extent. Furthermore, after the protein is modeled into a hypergraph, the hypergraph neural network can continue to be used for feature extraction, and the protein embedding vector can be well obtained for learning downstream tasks. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] In order to more clearly illustrate the technical methods in the embodiments of the present application, the drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For those skilled in the art, other drawings can be obtained based on these drawings without paying any creative work.

[0017] Figure 1 A schematic diagram of a construction method provided in an embodiment of the present application;

[0018] Figure 2 An example diagram of a construction method provided in an embodiment of the present application;

[0019] Figure 3 Another schematic diagram of the construction method provided in the embodiment of the present application;

[0020] Figure 4Another schematic diagram of the construction method provided in the embodiment of the present application;

[0021] Figure 5 Another example diagram of the construction method provided in the embodiment of the present application;

[0022] Figure 6 Another schematic diagram of the construction method provided in the embodiment of the present application;

[0023] Figure 7 Another example diagram of the construction method provided in the embodiment of the present application;

[0024] Figure 8 Another schematic diagram of the construction method provided in the embodiment of the present application;

[0025] Fig. 9 A schematic diagram of the structure of the construction device provided in the embodiment of the present application;

[0026] Fig.10 A schematic structural diagram of a computer device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0027] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical methods in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative work are within the scope of protection of this application.

[0028] First, some nouns or terms that appear in the description of the embodiments of the present application are explained as follows:

[0029] Blockchain system: It can be a distributed system formed by connecting clients and multiple nodes (any form of computing devices connected to the network, such as servers and user terminals) through network communication. The nodes form a peer-to-peer (P2P) network. The P2P protocol is an application layer protocol running on the Transmission Control Protocol (TCP). In a distributed system, any machine such as a server or terminal can join and become a node. The node includes the hardware layer, the middle layer, the operating system layer and the application layer.

[0030] Protein: It is an important component of all cells and tissues in the human body. Generally speaking, protein accounts for about 18% of the total mass of the human body. The most important thing is that it is related to life phenomena. There are many types of proteins in the human body, with different properties and functions, but they are all composed of 20 kinds of amino acids in different proportions, and are constantly metabolized and renewed in the body. Protein is a biological macromolecule composed of one or more polypeptide chains, each of which has twenty to hundreds of amino acid residues; various amino acid residues are arranged in a certain order. The organelle that produces protein is the ribosome.

[0031] Aminoacid: It is the basic unit of protein, which gives protein a specific molecular structure and makes its molecule biochemically active. Protein is an important active molecule in organisms, including enzymes and enzymes that catalyze metabolism. Different amino acids are chemically polymerized into peptides, a primitive fragment of protein, which is the precursor of protein production.

[0032] Hypergraph: A hypergraph is a generalization of a graph. Unlike an edge in a graph (an edge only connects two vertices), a hyperedge in a hypergraph can connect multiple vertices.

[0033] The current representation of proteins is based on sequences or simple graphs. The corresponding learning of proteins is to use sequence models or graph models in deep learning. Sequence models use chains of amino acids in proteins as representations of proteins, while graph models generally refer to modeling amino acids as vertices of a graph, and two amino acids within a certain distance are considered to have an edge.

[0034] The embodiments of the present application provide a method for constructing a protein hypergraph, a device for constructing a protein hypergraph, a medium, an equipment and a program product. The method for constructing a protein hypergraph of the present application constructs multiple amino acid interaction sequence structure hyperedges and spatial structure hyperedges in space, and also incorporates the characteristic information of vertex amino acids, so that the constructed hypergraph can reflect the spatial relationship between amino acids and the characteristics of amino acids, and can model the high-order information of proteins to a certain extent. The embodiments of the present application use hypergraphs to model the high-order information of proteins, which can be applied to the field of protein representation learning. At the same time, on the one hand, it can also be applied to protein classification, which is very important for predicting protein function and is of great significance in the field of life sciences. On the other hand, it can be used to score the predicted spatial structure of existing proteins and estimate the degree of fit with the real protein. After the model is learned, it can be used for quality assessment after protein generation.

[0035] Specifically, the method of the embodiment of the present application may be executed by a computer device, wherein the computer device may be a terminal or a server.

[0036] In order to better understand the technical method provided in the embodiment of the present application, the following briefly introduces the application scenarios applicable to the technical method provided in the embodiment of the present application. It should be noted that the application scenarios introduced below are only used to illustrate the embodiment of the present application and are not limited. Take the construction method executed by a computer device as an example, where the computer device can be a terminal or a server.

[0037] The embodiments of the present application can be implemented in combination with cloud technology or blockchain network technology. As disclosed in the construction method of the embodiments of the present application, these data can be stored on the blockchain, for example: structural hyperedges, first hyperedge sets, amino acid features, vertex feature vectors, vertex feature sets and protein hypergraphs can all be stored on the blockchain.

[0038] In order to facilitate the storage and query of the structural hyperedges, the first hyperedge set, the amino acid features, the vertex feature vectors, the vertex feature set and the protein hypergraph, optionally, the construction method also includes: sending the structural hyperedges, the first hyperedge set, the amino acid features, the vertex feature vectors, the vertex feature set and the protein hypergraph to the blockchain network, so that the nodes of the blockchain network fill the structural hyperedges, the first hyperedge set, the amino acid features, the vertex feature vectors, the vertex feature set and the protein hypergraph into the new block, and when a consensus is reached on the new block, the new block is appended to the end of the blockchain. The embodiment of the present application can store the structural hyperedges, the first hyperedge set, the amino acid features, the vertex feature vectors, the vertex feature set and the protein hypergraph on the chain to realize the backup of the records. When the protein hypergraph needs to be obtained, the corresponding protein hypergraph can be directly and quickly obtained from the blockchain, thereby improving the efficiency of the protein hypergraph construction method.

[0039] It should be noted that the order of description of the following embodiments is not intended to limit the priority order of the embodiments.

[0040] Each embodiment of the present application provides a method for constructing a protein hypergraph. The embodiments of the present application use a computer device as an example to illustrate the construction method.

[0041] See also Figure 1 , Figure 1 A schematic diagram of a method for constructing a protein hypergraph provided in an embodiment of the present application, the method comprising:

[0042] Step 110: construct multiple structural hyperedges according to the spatial relationship between all amino acids contained in the protein to obtain a first hyperedge set of the protein, wherein the structural hyperedges contain multiple amino acids.

[0043] Specifically, in the spatial structure of protein molecules, all proteins are polymers formed by the connection of 20 different amino acids. After forming proteins, these amino acids are called residues. Amino acids are the basic units that make up proteins, and they form different proteins in different orders and configurations.

[0044] A hypergraph is a generalization of a graph. Unlike an edge in a graph that only connects two vertices, a hyperedge in a hypergraph can connect two or more vertices. There are many ways to construct a hyperedge, which can represent the relationship between multiple connected vertices to a certain extent. Figure 2 ,The figure is an example graph of hyperedges, where the nodes are n1-n8 for a total of 8 nodes, e1, e2 and e3 are 3 hyperedges, for example, the hyperedge e1 connects the nodes n2, n4 and n8.

[0045] Among them, the spatial relationship between amino acids includes the distance relationship between amino acids, the spatial relationship between adjacent amino acids, etc.

[0046] Optionally, the structural hyperedge includes a spatial structural hyperedge, and the step of constructing a plurality of structural hyperedges according to the spatial relationship between all amino acids contained in the protein to obtain a first hyperedge set of the protein includes:

[0047] Get the current amino acid and calculate the Euclidean distance between the current amino acid and any target amino acid in the protein;

[0048] When the Euclidean distance is within the distance threshold, the current amino acid and the target amino acid are determined as the first vertex amino acids;

[0049] Determine a spatial structure hyperedge according to a plurality of first vertex amino acids;

[0050] According to the spatial structure hyperedges determined by all the amino acids of the protein, a first hyperedge set is determined.

[0051] Specifically, a protein molecule is obtained, a current amino acid is determined, and other amino acids in the protein are used as target amino acids. The Euclidean distance between the current amino acid and any target amino acid is calculated and compared with a distance threshold, which is a preset threshold, such as 4A. 0 , 6A 0 .

[0052] The target amino acid that meets this condition may be one or more amino acids: when the calculated Euclidean distance is within the distance threshold range, or in other words, the calculated Euclidean distance is less than or equal to the distance threshold, the target amino acid and the current amino acid may be determined as the first vertex amino acid. At the same time, these target amino acids and the current amino acid determine a spatial structure hyperedge, or in other words, the spatial structure hyperedge includes the target amino acid and the current amino acid whose Euclidean distance is within the distance threshold range.

[0053] Traverse each amino acid in the protein, take each amino acid as the current amino acid, calculate the Euclidean distance with any other target amino acid, and obtain the spatial structure hyperedge of each amino acid. When all amino acids have determined one or more spatial structure hyperedges, determine the first hyperedge set based on the spatial structure hyperedges of all amino acids in the protein. The first hyperedge set includes all spatial structure hyperedges of all amino acids and the amino acids contained in each hyperedge.

[0054] The spatial structure hyperedge can be expressed as:

[0055] e i = {v j |d(v i , v j )<d cut_off};

[0056] Among them, e i Represented as the hyperedge of the ith amino acid, v j represents the jth amino acid, d(v i , v j ) represents the Euclidean distance between the ith amino acid and the jth amino acid, d cut_off Indicates the limited distance threshold, d cut_off >0.

[0057] Optionally, the structural hyperedge includes a sequence structural hyperedge, and the step of constructing a plurality of structural hyperedges according to the spatial relationship between all amino acids contained in the protein to obtain a first hyperedge set of the protein includes:

[0058] Determine the adjacent k amino acids as the second vertex amino acids, and connect them to obtain a sequence structure hyperedge, where k is an integer;

[0059] A first hyperedge set is determined according to all sequence structure hyperedges in the protein and / or spatial structure hyperedges determined by all amino acids.

[0060] Specifically, k adjacent amino acids can be connected as a sequence structure hyperedge, and k amino acids can be used as second vertex amino acids. For example, five adjacent amino acids can be determined as second vertex amino acids to form a sequence structure hyperedge.

[0061] The i-th sequence structure hyperedge can be expressed as: i = {v i+1 , v i+2 ,…v i+k}, where v i+1 It is represented by the i+1th amino acid, i.e. e i It is represented as the i-th sequence structure hyperedge formed by connecting k adjacent amino acids together.

[0062] The first hyperedge set may include sequence structure hyperedges, or space structure hyperedges, or the sequence structure hyperedges and the space structure hyperedges together form the first hyperedge set.

[0063] Step 120: Obtaining the amino acid feature of each amino acid, and constructing a vertex feature vector of each amino acid according to the amino acid feature to obtain a vertex feature set of the protein;

[0064] Optionally, the vertex feature vector includes a first vertex feature vector, and the step of obtaining the amino acid feature of each amino acid and constructing the vertex feature vector of each amino acid according to the amino acid feature to obtain the vertex feature set of the protein includes:

[0065] Obtain the secondary structure of all amino acids in a protein;

[0066] A first real number vector is generated according to the secondary structure of each amino acid to obtain a first vertex feature vector of the protein. The first real number vector is used to characterize the secondary structure of the amino acid.

[0067] Specifically, amino acids have 3D spatial structural information, such as the secondary structure of amino acids. The secondary structure can be obtained from the pdb file storing protein data through the existing Python toolkit DSSP, wherein the method used is Bio.PDB.DSSP.dssp_dict_from_pdb_file().

[0068] The secondary structure of a protein is the structure of each amino acid in the amino acid chain, which is mainly divided into eight categories: "G", "H", "I", "T", "E", "B", "S", "Blank". Among them, H = α-helix, B = β-bridge, E = β-strand, G = 3helix (3 / 10helix), I = 5helix (pihelix), T = β-turn, S = bend, Blank is a null value.

[0069] For example, a 7-dimensional first real vector is generated according to the secondary structure of each amino acid, including generating a one-hot vector, to obtain the first vertex feature vector of the protein, and the first real vector is used to characterize the secondary structure of the amino acid. For example, a common secondary structure is HBET, and the first real vector represented in one-hot form is {1, 1, 1, 0, 0, 1, 0, 0}.

[0070] Optionally, the vertex feature includes a second vertex feature vector, and the step of obtaining the amino acid feature of each amino acid and constructing the vertex feature vector of each amino acid according to the amino acid feature to obtain the vertex feature set of the protein also includes:

[0071] Obtaining spatial characteristic information of the protein, the spatial characteristic information includes the relative accessible surface area (rASA) of the protein residues, the first torsion angle Phi, and the second torsion angle Psi;

[0072] A second real number vector is generated according to the spatial feature information to obtain a second vertex feature vector of the protein.

[0073] Specifically, the spatial characteristic information of a protein includes the relative accessible surface area (rASA) of protein residues, the first torsion angle Phi, and the second torsion angle Psi.

[0074] Among them, accessible surface area (ASA) or solvent accessible surface area (SASA) is the surface area of ​​a biomacromolecule that is accessible to solvent. The relative accessible surface area (rASA) or relative solvent accessibility of protein residues is a way to measure the solvent exposure of residues. It can be calculated by the formula.

[0075] rASA=ASA / MaxASA;

[0076] Where ASA is the solvent accessible surface area and MaxASA is the maximum possible solvent accessible surface area of ​​the residue.

[0077] The four atoms that make up the first torsion angle Phi are a carbonyl carbon, the attached α-carbon, an amide nitrogen, and the next carbonyl carbon (all marked with green halos). The four atoms that make up the second torsion angle Psi are an amide nitrogen, a carbonyl carbon, an α-carbon, and the second nitrogen.

[0078] It can be understood that the tetrapeptide Leu-Leu-Ile-Tyr may contain four amino acids linked together by three amide bonds or peptide bonds. Since the amide bond has the characteristics of a π bond, the six atoms that make up the peptide bond are all located in the same plane, where the first plane is the peptide bond connecting Tyr and Ile, and the second plane connects Ile and Leu. As with any peptide, the conformation of the backbone is determined by the values ​​of two torsion angles. In order, phi (φ) is the C(i-), N(i), Ca(i), C(i) torsion angle, and psi (ψ) is the N(i), Ca(i), C(i), N(i+1) torsion angle. Then, the second plane is used as a reference for measuring the two angles. Then the third plane is part of the planar peptide bond between Tyr and Ile (the fourth plane), and the angle between the fourth plane and the second plane is psi. The first plane is part of the planar peptide bond between Ile and Leu (the fifth plane), and the angle between the fifth plane and the second plane is phi.

[0079] Calculate the values ​​of the spatial feature information rASA, Phi and Psi of the protein, and generate a second real number vector according to the values ​​of rASA, Phi and Psi to obtain a 3D second vertex feature vector of the protein. For example, the second vertex feature vector is {1.1000e+02, 3.6000e+02, 1.3190e+02}.

[0080] Optionally, the vertex feature includes a third vertex feature vector, and the step of obtaining the amino acid feature of each amino acid and constructing the vertex feature vector of each amino acid according to the amino acid feature to obtain the vertex feature set of the protein also includes:

[0081] A third real number vector is generated according to all the amino acids of the protein to obtain a third vertex feature vector of the protein. The third real number vector is used to characterize the types of amino acids contained in the protein.

[0082] Specifically, the main amino acid types that make up the proteins in living organisms include 20 amino acids: glycine, alanine, valine, leucine, isoleucine, methionine (methionine), proline, tryptophan, serine, tyrosine, cysteine, phenylalanine, asparagine, glutamine, threonine, aspartic acid, glutamic acid, lysine, arginine and histidine. A third real number vector is generated according to the types of all amino acids contained in the protein to obtain the third vertex eigenvector of the protein.

[0083] For example, 20 common amino acids are represented using one-hot encoding as follows:

[0084] 'A': [1,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0]

[0085] ‘C’:[0,1,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0]

[0086] ‘D’:[0,0,1,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0]

[0087] ‘E’:[0,0,0,1,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0]

[0088] ‘F’:[0,0,0,0,1,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0]

[0089] ‘G’:[0,0,0,0,0,1,0,0,0,0,0,0,0,0,0,0,0,0,0,0]

[0090] ‘H’:[0,0,0,0,0,0,1,0,0,0,0,0,0,0,0,0,0,0,0,0]

[0091] ‘I’:[0,0,0,0,0,0,0,1,0,0,0,0,0,0,0,0,0,0,0,0]

[0092] ‘K’:[0,0,0,0,0,0,0,0,1,0,0,0,0,0,0,0,0,0,0,0]

[0093] ‘L’:[0,0,0,0,0,0,0,0,0,1,0,0,0,0,0,0,0,0,0,0]

[0094] ‘M’:[0,0,0,0,0,0,0,0,0,0,1,0,0,0,0,0,0,0,0,0]

[0095] ‘N’:[0,0,0,0,0,0,0,0,0,0,0,1,0,0,0,0,0,0,0,0]

[0096] ‘P’:[0,0,0,0,0,0,0,0,0,0,0,0,1,0,0,0,0,0,0,0]

[0097] ‘Q’:[0,0,0,0,0,0,0,0,0,0,0,0,0,1,0,0,0,0,0,0]

[0098] ‘R’:[0,0,0,0,0,0,0,0,0,0,0,0,0,0,1,0,0,0,0,0]

[0099] 'S': [0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,1,0,0,0,0]

[0100] 'T': [0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,1,0,0,0]

[0101] 'V': [0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,1,0,0]

[0102] 'W': [0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,1,0]

[0103] 'Y': [0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,1]

[0104] The vertex feature vector of the protein may include the first feature vector, the second feature vector, and the third feature vector connected together, and all feature vectors together, then each vertex has a 29-dimensional vector. The dimension of the vertex feature matrix X is |V|×29, where V represents amino acids.

[0105] Optionally, the vertex feature includes a fourth vertex feature vector, and the method further includes:

[0106] Based on the structure of each amino acid in the protein, the statistical characteristics of chemical elements are obtained;

[0107] The fourth vertex feature vector is generated according to the chemical element statistical characteristics of each amino acid.

[0108] Specifically, according to the structure of each amino acid of the protein, the chemical element statistical characteristics are obtained, wherein the chemical element statistical characteristics may include the relationship between the chemical elements contained in the amino acid, such as the number of benzene rings and the number of hydroxyl groups in the amino acid, etc. The chemical element statistical characteristics of each amino acid are aggregated to generate the fourth vertex feature vector.

[0109] Step 130: Determine a protein hypergraph according to the first hyperedge set and the vertex feature set.

[0110] Specifically, when the first hyperedge set and vertex feature set are determined, a protein hypergraph can be obtained. For example, the protein hypergraph H(V, E1) is composed of the vertex feature set V and the first hyperedge set E1. Protein hypergraph is a method to represent proteins through multimodal data. The interaction relationship between amino acids, that is, the high-order information in the protein, can be represented through hyperedge and vertex features.

[0111] After modeling the protein into a hypergraph, the hypergraph neural network can be used for feature extraction to obtain the protein embedding vector for learning downstream tasks, such as protein classification, protein quality assessment, etc.

[0112] In this way, the present application constructs multiple structural hyperedges according to the spatial relationship between all the amino acids contained in the protein to obtain the first hyperedge set of the protein, and the structural hyperedges contain multiple amino acids; obtains the amino acid features of each amino acid, and constructs the vertex feature vector of each amino acid according to the amino acid features to obtain the vertex feature set of the protein; determines the protein hypergraph according to the first hyperedge set and the vertex feature set. The protein hypergraph construction method of the present application constructs multiple amino acid interaction sequence structural hyperedges and spatial structural hyperedges in space, and also incorporates the feature information of the vertex amino acids, so that the constructed hypergraph can reflect the spatial relationship and amino acid features between amino acids, and can model the high-order information of the protein to a certain extent. Furthermore, after the protein is modeled into a hypergraph, the hypergraph neural network can continue to be used for feature extraction, and the protein embedding vector can be well obtained for learning downstream tasks.

[0113] See also Figure 3 , FIG. 1 is a flow chart of another method for constructing a protein supergraph provided in an embodiment of the present application, the method comprising:

[0114] Step 310, constructing a plurality of structural hyperedges according to the spatial relationship between all amino acids contained in the protein to obtain a first hyperedge set of the protein, wherein the structural hyperedges include a plurality of amino acids;

[0115] The specific implementation is the same as the above step 110 and will not be described in detail here.

[0116] Step 320, obtaining the amino acid feature of each amino acid, and constructing the vertex feature vector of each amino acid according to the amino acid feature to obtain the vertex feature set of the protein;

[0117] The specific implementation is the same as the above step 120 and will not be described in detail here.

[0118] Step 330, constructing a feature hyperedge according to the distance between the vertex feature vectors of each amino acid;

[0119] Specifically, the vertex feature vectors obtained in any of the above embodiments include a first vertex feature vector, a second vertex feature vector, a third vertex feature vector and a fourth vertex feature vector. The distance between the vertex feature vectors of each amino acid, such as the Euclidean distance, is calculated, and the distance is compared with a predetermined distance threshold. The amino acids within the distance threshold range are determined as vertices in a feature hyperedge, thereby constructing a feature hyperedge.

[0120] Step 340, determining a second hyperedge set based on characteristic hyperedges determined by all amino acids of the protein;

[0121] Step 350 , determining a protein hypergraph according to the first hyperedge set, the second hyperedge set, and the vertex feature set.

[0122] For example, the characteristic hyperedges of each amino acid in the protein are collected to obtain the second hyperedge set of the protein. The protein hypergraph is determined based on the first hyperedge set, the second hyperedge set and the vertex feature set. For example, the protein hypergraph H(V, E1, E2) is composed of the vertex feature set V, the first hyperedge set E1 and the second hyperedge set E2.

[0123] In this way, extending the hyperedge design of the protein hypergraph to the vertex feature vectors can obtain richer high-order information of the protein and improve the efficiency of protein representation.

[0124] See also Figure 4 , FIG. 1 is a flow chart of another method for constructing a protein supergraph provided in an embodiment of the present application, the method comprising:

[0125] Step 410, constructing a plurality of structural hyperedges according to the spatial relationship between all amino acids contained in the protein to obtain a first hyperedge set of the protein, wherein the structural hyperedges contain a plurality of amino acids;

[0126] The specific implementation is the same as the above step 110 and will not be described in detail here.

[0127] Step 420, obtaining the amino acid feature of each amino acid, and constructing the vertex feature vector of each amino acid according to the amino acid feature to obtain the vertex feature set of the protein;

[0128] The specific implementation is the same as the above step 120 and will not be described in detail here.

[0129] Step 430, constructing a feature hyperedge according to the distance between the vertex feature vectors of each amino acid;

[0130] Step 440, determining a second hyperedge set based on characteristic hyperedges determined by all amino acids of the protein;

[0131] Step 450, using a kernel function to perform spatial transformation on the Euclidean distance to obtain a vertex weight of each first vertex amino acid in the spatial structure hyperedge, where the vertex weight is used to characterize the connection strength between the first vertex amino acid and the spatial structure hyperedge;

[0132] Specifically, the distance is spatially transformed using a kernel function to obtain the vertex weight of each first vertex amino acid in the spatial structure hyperedge. The vertex weight Q 1 (c, j) can be obtained by the following formula:

[0133]

[0134] Among them, d(v,v c ) 2 Represents vertices v and v c The Euclidean distance. j represents the jth hyperedge, each hyperedge is represented by multiple vertices, such as the sequence structure hyperedge e i = {v i+1 , v i+2 ,…v i+k}, or spatial structure hyperedge: e i = {v j |d(v i , v j )<d cut_off};

[0135] Represents vertex v c The average distance to all other vertices, It can be expressed as:

[0136]

[0137] Where V represents the number of vertices.

[0138] Step 460, determining the spatial structure association matrix of the protein according to the vertex weight corresponding to each first vertex amino acid;

[0139] After obtaining the vertex weight corresponding to each first vertex amino acid, the spatial structure association matrix Q∈R of all amino acids is formed. |V|×|E| , Q is a VxE dimensional matrix.

[0140] Step 470, determining a protein hypergraph according to the first hyperedge set, the second hyperedge set, the vertex feature set, and the spatial structure association matrix.

[0141] For example, the protein hypergraph H(V, E, Q1) mainly consists of the vertex feature set V, the hyperedge set E consisting of the first hyperedge set E1 and the second hyperedge set E2, and the spatial structure association matrix Q1.

[0142] See also Figure 5 , the figure is a representation of the spatial structure association matrix, Q is the spatial structure association matrix, e1 is the spatial structure association hyperedge 1, e2 is the spatial structure association hyperedge 2, a, b, c, d and f are vertex amino acids, among which hyperedge e1 includes vertex amino acids a, b, c, d, hyperedge e2 includes vertex amino acids b, d and f, Q(a), Q(b), Q(c), Q(d) and Q(f) are the vertex weights of the vertices.

[0143] In this way, the spatial structure association matrix is ​​added to the protein hypergraph and the vertex weights of amino acids are introduced, so that the constructed hypergraph can reflect the spatial relationship between amino acids.

[0144] See also Figure 6 , FIG. 1 is a flow chart of another method for constructing a protein supergraph provided in an embodiment of the present application, the method comprising:

[0145] Step 610, constructing a plurality of structural hyperedges according to the spatial relationship between all amino acids contained in the protein to obtain a first hyperedge set of the protein, wherein the structural hyperedges include a plurality of amino acids;

[0146] The specific implementation is the same as the above step 11θ, and will not be described in detail here.

[0147] Step 620, obtaining the amino acid feature of each amino acid, and constructing the vertex feature vector of each amino acid according to the amino acid feature to obtain the vertex feature set of the protein;

[0148] The specific implementation is the same as the above step 120 and will not be described in detail here.

[0149] Step 630, constructing characteristic hyperedges according to the distances between vertex feature vectors of each amino acid, and determining a second hyperedge set according to the characteristic hyperedges determined by all amino acids of the protein;

[0150] The specific implementation is the same as the above embodiment and will not be described in detail here.

[0151] Step 640, using a kernel function to perform spatial transformation on the Euclidean distance to obtain a vertex weight of each first vertex amino acid in the spatial structure hyperedge, and determining the spatial structure association matrix of the protein according to the vertex weight corresponding to each first vertex amino acid;

[0152] The specific implementation is the same as the above embodiment and will not be described in detail here.

[0153] Step 650, determining a sequence structure association matrix according to a plurality of sequence structure hyperedges;

[0154] Step 660, concatenating the sequence structure association matrix and the spatial structure association matrix to obtain a structure matrix of the protein;

[0155] All sequence structure hyperedges of V amino acids in the protein are collected to obtain a VxV-dimensional sequence structure association matrix Q2. The spatial structure association matrix Q1 is concatenated with the sequence structure association matrix Q2 to obtain a structure matrix Q = [Q1, Q2].

[0156] Step 670, determining a protein hypergraph according to the first hyperedge set, the second hyperedge set, the vertex feature set and the structure matrix.

[0157] The protein hypergraph H(V, E, Q) consists of a vertex feature set V, a hyperedge set E consisting of a first hyperedge set E1 and a second hyperedge set E2, and a structure matrix Q = [Q1, Q2].

[0158] See also Figure 7 , Figure 7 The ordinate represents the number of amino acids, the abscissa represents the number of hyperedges, the left frame of the figure is the spatial structure hyperedge Q1, the right frame is the sequence structure hyperedge Q2, the white dot part of Q1 in the figure represents the position of the vertex weight Q1, and the white line part of Q2 represents the position of the sequence structure association matrix Q2.

[0159] See also Figure 8 , FIG. 1 is a flow chart of another method for constructing a protein supergraph provided in an embodiment of the present application, the method comprising:

[0160] Step 810, constructing a plurality of structural hyperedges according to the spatial relationship between all amino acids contained in the protein to obtain a first hyperedge set of the protein, wherein the structural hyperedges include a plurality of amino acids;

[0161] The specific implementation is the same as the above step 110 and will not be described in detail here.

[0162] Step 820, obtaining the amino acid feature of each amino acid, and constructing the vertex feature vector of each amino acid according to the amino acid feature to obtain the vertex feature set of the protein;

[0163] The specific implementation is the same as the above step 120 and will not be described in detail here.

[0164] Step 830, constructing characteristic hyperedges according to the distances between vertex feature vectors of each amino acid, and determining a second hyperedge set according to the characteristic hyperedges determined by all amino acids of the protein;

[0165] The specific implementation is the same as the above embodiment and will not be described in detail here.

[0166] Step 840, using a kernel function to perform spatial transformation on the Euclidean distance to obtain a vertex weight of each first vertex amino acid in the spatial structure hyperedge, and determining the spatial structure association matrix of the protein according to the vertex weight corresponding to each first vertex amino acid;

[0167] The specific implementation is the same as the above embodiment and will not be described in detail here.

[0168] Step 850, determining a sequence-structure association matrix according to the plurality of sequence-structure hyperedges, and concatenating the sequence-structure association matrix and the spatial structure association matrix to obtain a structure matrix of the protein;

[0169] Step 860, generating a corresponding edge weight diagonal matrix according to the first super edge set or the second super edge set, where the edge weight diagonal matrix is ​​a unit matrix;

[0170] Generate the corresponding edge weight diagonal matrix according to the first hyperedge set or the second hyperedge set. The edge weight diagonal matrix is ​​the unit matrix W, and W has the same dimension as the hyperedge set, for example, W∈R |E|×|E| .

[0171] Step 870, determining a protein hypergraph according to the first hyperedge set, the second hyperedge set, the vertex feature set, the structural association matrix, and the edge weight diagonal matrix;

[0172] The protein hypergraph H(V, E, Q, W) consists of the vertex feature set V, the hyperedge set E consisting of the first hyperedge set E1 and the second hyperedge set E2, the structure matrix Q = [Q1, Q2], and the edge weight diagonal matrix W∈R |E|×|E| composition.

[0173] Step 880, sending the protein hypergraph to the hypergraph neural network to perform representation learning on the protein.

[0174] Specifically, Hypergraph Neural Networks (HGNN) use existing hyperedge convolution operations to handle data correlation in the representation learning process. Through hypergraph neural networks, hyperedge convolution operations can be effectively used to perform traditional protein hypergraph learning. HGNN is able to learn hidden layer representations that take into account high-order data structures. HGNN first takes multimodal data relations, namely protein hypergraphs, as input, considers them simultaneously for hyperedge generation, and performs hypergraph convolution on the generated hyperedge groups to obtain the final result, which is output according to the requirements of downstream tasks.

[0175] It should be noted that the edge weight diagonal matrix is ​​the unit matrix added according to HGNN. When the protein hypergraph is sent to different networks, the protein hypergraph can be adaptively adjusted to the input format of the network according to the needs of the network.

[0176] All of the above technical methods can be combined in any way to form optional embodiments of the present application, and will not be described in detail here.

[0177] In order to better implement the protein hypergraph construction method of the present application embodiment, the present application embodiment also provides a protein hypergraph construction device. Fig. 9 , Fig. 9 A schematic diagram of the structure of a protein hypergraph construction device provided in an embodiment of the present application. The construction device 900 may include:

[0178] A first construction unit 910 is used to construct a plurality of structural hyperedges according to a plurality of amino acids of the protein to obtain a first hyperedge set of the protein, wherein the structural hyperedges include a plurality of vertex amino acids;

[0179] The second construction unit 920 is used to obtain the amino acid features of each vertex amino acid, and construct a vertex feature vector of each vertex amino acid according to the amino acid features to obtain a vertex feature set of the protein;

[0180] The determining unit 930 is used to determine the protein hypergraph according to the first hyperedge set and the vertex feature set.

[0181] Optionally, the constructing device 900 may further include:

[0182] The third construction unit 940 is used to construct a characteristic hyperedge according to the distance between the vertex feature vectors of each amino acid; determine the second hyperedge set based on the characteristic hyperedges determined by all amino acids of the protein; the determination unit 930 can also be used to determine the protein hypergraph based on the first hyperedge set, the second hyperedge set and the vertex feature set.

[0183] Optionally, the constructing device 900 may further include:

[0184] The transformation unit 950 is used to perform spatial transformation on the Euclidean distance using a kernel function to obtain a vertex weight of each first vertex amino acid in the spatial structure hyperedge, wherein the vertex weight is used to characterize the connection strength between the first vertex amino acid and the spatial structure hyperedge; and determine the spatial structure association matrix of the protein according to the vertex weight corresponding to each first vertex amino acid;

[0185] Accordingly, the determining unit 930 is used to determine the protein hypergraph according to the first hyperedge set, the second hyperedge set, the vertex feature set and the spatial structure association matrix.

[0186] Optionally, the constructing device 900 may further include:

[0187] A concatenation unit 960 is used to determine a sequence-structure association matrix according to a plurality of sequence-structure hyperedges; and concatenate the sequence-structure association matrix and the spatial structure association matrix to obtain a structure matrix of the protein;

[0188] Accordingly, the determining unit 930 is used to determine the protein hypergraph according to the first hyperedge set, the second hyperedge set, the vertex feature set and the structure matrix.

[0189] Optionally, the constructing device 900 may further include:

[0190] The generation unit 970 is used to generate a corresponding edge weight diagonal matrix according to the first hyperedge set or the second hyperedge set, where the edge weight diagonal matrix is ​​a unit matrix; determine a protein hypergraph according to the first hyperedge set, the second hyperedge set, the vertex feature set, the structural association matrix and the edge weight diagonal matrix; and send the protein hypergraph to a hypergraph neural network to perform representation learning on the protein.

[0191] Optionally, the first construction unit 910 can also be used to obtain the current amino acid and calculate the Euclidean distance between the current amino acid and any target amino acid in the protein; when the Euclidean distance is within a distance threshold range, the current amino acid and the target amino acid are determined as first vertex amino acids; a spatial structure hyperedge is determined based on multiple first vertex amino acids; and a first hyperedge set is determined based on the spatial structure hyperedges determined by all amino acids of the protein.

[0192] Optionally, the first construction unit 910 can also be used to determine k adjacent amino acids as second vertex amino acids and connect them to obtain sequence structure hyperedges, where k is an integer; and determine the first hyperedge set based on all sequence structure hyperedges in the protein and / or the spatial structure hyperedges determined by all amino acids.

[0193] Optionally, the second construction unit 920 can also be used to obtain the secondary structure of all amino acids in the protein; generate a first real number vector according to the secondary structure of each amino acid to obtain a first vertex feature vector of the protein, and the first real number vector is used to characterize the secondary structure of the amino acid.

[0194] Optionally, the second construction unit 920 can also be used to obtain spatial feature information of the protein, the spatial feature information including the relative accessible surface area (rASA) of the protein residues, the first torsion angle Phi, and the second torsion angle Psi; generate a second real number vector based on the spatial feature information to obtain a second vertex feature vector of the protein.

[0195] Optionally, the second construction unit 920 may also be used to generate a third real number vector according to all the amino acids of the protein to obtain a third vertex feature vector of the protein, where the third real number vector is used to characterize the types of amino acids contained in the protein.

[0196] Optionally, the second construction unit 920 may also be used to obtain chemical element statistical characteristics according to the structure of each amino acid of the protein; and generate a fourth vertex feature vector according to the chemical element statistical characteristics of each amino acid.

[0197] It should be noted that the functions of each module in the construction device 900 in the embodiment of the present application can correspond to the specific implementation method of any embodiment in the above-mentioned method embodiments, and will not be repeated here.

[0198] Each unit in the construction device 900 can be implemented in whole or in part by software, hardware, or a combination thereof. Each unit can be embedded in or independent of a processor in a computer device in the form of hardware, or can be stored in a memory in a computer device in the form of software, so that the processor can call and execute operations corresponding to each unit.

[0199] The construction device 900 can be integrated in a terminal or server with a storage device and a processor installed and having computing power, or the construction device 900 is the terminal or server. The terminal can be a smart phone, a tablet computer, a laptop computer, a smart TV, a smart speaker, a wearable smart device, a personal computer (PC) and other devices, and the terminal can also include a client, which can be a video client, a browser client or an instant messaging client. The server can be an independent physical server, or a server cluster or distributed system composed of multiple physical servers, or a cloud server that provides basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communications, middleware services, domain name services, security services, content delivery networks (CDNs), and big data and artificial intelligence platforms.

[0200] Fig.10 A schematic structural diagram of a computer device 1000 provided in an embodiment of the present application is shown in the figure. The computer device may include: a communication interface 1001, a memory 1002, a processor 1003 and a communication bus 1004. The communication interface 1001, the memory 1002, and the processor 1003 communicate with each other through the communication bus 1004. The communication interface 1001 is used for the device 1000 to communicate data with an external device. The memory 1002 can be used to store software programs and modules, and the processor 1003 runs the software programs and modules stored in the memory 1002, such as the software programs of the corresponding operations in the aforementioned method embodiment.

[0201] Optionally, the processor 1003 may call the software program and module stored in the memory 1002 to perform the following operations:

[0202] Constructing multiple structural hyperedges according to the spatial relationship between all amino acids contained in the protein to obtain a first hyperedge set of the protein, wherein the structural hyperedges contain multiple amino acids;

[0203] Obtaining the amino acid features of each amino acid, and constructing the vertex feature vector of each amino acid according to the amino acid features to obtain the vertex feature set of the protein;

[0204] A protein hypergraph is determined according to the first hyperedge set and the vertex feature set.

[0205] Optionally, the computer device is the terminal or server. The terminal may be a smart phone, tablet computer, laptop computer, smart TV, smart speaker, wearable smart device, personal computer, game console, vehicle terminal, smart TV and other devices. The server may be an independent physical server, or a server cluster or distributed system composed of multiple physical servers, or a cloud server that provides basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communications, middleware services, domain name services, security services, CDN, and big data and artificial intelligence platforms.

[0206] The present application also provides a computer-readable storage medium for storing a computer program. The computer-readable storage medium can be applied to a computer device, and the computer program enables the computer device to execute the corresponding process in the method for constructing a protein hypergraph in the embodiment of the present application, which will not be described in detail here for the sake of brevity.

[0207] The present application also provides a computer program product, which includes computer instructions, which are stored in a computer-readable storage medium. The processor of the computer device reads the computer instructions from the computer-readable storage medium, and the processor executes the computer instructions, so that the computer device executes the corresponding process in the method for constructing a protein hypergraph in the embodiment of the present application, which will not be described in detail here for the sake of brevity.

[0208] The present application also provides a computer program, which includes computer instructions, and the computer instructions are stored in a computer-readable storage medium. The processor of the computer device reads the computer instructions from the computer-readable storage medium, and the processor executes the computer instructions, so that the computer device executes the corresponding process in the method for constructing a protein hypergraph in the embodiment of the present application, which will not be described here for the sake of brevity.

[0209] It should be understood that the processor of the embodiment of the present application may be an integrated circuit chip with signal processing capabilities. In the implementation process, each step of the above method embodiment can be completed by the hardware integrated logic circuit or software instructions in the processor. The above processor can be a general processor, a digital signal processor (Digital Signal Processor, DSP), an application-specific integrated circuit (Application Specific Integrated Circuit, ASIC), a field programmable gate array (Field Programmable Gate Array, FPGA) or other programmable logic devices, discrete gates or transistor logic devices, discrete hardware components. The methods, steps and logic block diagrams disclosed in the embodiments of the present application can be implemented or executed. The general processor can be a microprocessor or the processor can also be any conventional processor. The steps of the method disclosed in the embodiment of the present application can be directly embodied as a hardware decoding processor to perform, or the hardware and software modules in the decoding processor are combined and performed. The software module can be located in a random access memory, a flash memory, a read-only memory, a programmable read-only memory or an electrically erasable programmable memory, a register, and other mature storage media in the art. The storage medium is located in the memory, and the processor reads the information in the memory and completes the steps of the above method in combination with its hardware.

[0210] It can be understood that the memory in the embodiments of the present application can be a volatile memory or a non-volatile memory, or can include both volatile and non-volatile memories. Among them, the non-volatile memory can be a read-only memory (ROM), a programmable read-only memory (PROM), an erasable programmable read-only memory (EPROM), an electrically erasable programmable read-only memory (EEPROM), or a flash memory. The volatile memory can be a random access memory (RAM), which is used as an external cache. By way of example and not limitation, many forms of RAM are available, such as static random access memory (SRAM), dynamic random access memory (DRAM), synchronous dynamic random access memory (SDRAM), double data rate synchronous dynamic random access memory (DDR SDRAM), enhanced synchronous dynamic random access memory (ESDRAM), synchronous link dynamic random access memory (SLDRAM), and direct RAM bus random access memory (DR RAM). It should be noted that the memory of the systems and methods described herein is intended to include, but is not limited to, these and any other suitable types of memory.

[0211] It should be understood that the above-mentioned memory is exemplary but not restrictive. For example, the memory in the embodiments of the present application may also be static random access memory (static RAM, SRAM), dynamic random access memory (dynamic RAM, DRAM), synchronous dynamic random access memory (synchronous DRAM, SDRAM), double data rate synchronous dynamic random access memory (double data rate SDRAM, DDR SDRAM), enhanced synchronous dynamic random access memory (enhanced SDRAM, ESDRAM), synchronous link dynamic random access memory (synch link DRAM, SLDRAM) and direct memory bus random access memory (Direct Rambus RAM, DR RAM), etc. That is to say, the memory in the embodiments of the present application is intended to include but not limited to these and any other suitable types of memory.

[0212] Those of ordinary skill in the art will appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical method. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of this application.

[0213] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the systems, devices and units described above can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0214] In the several embodiments provided in the present application, it should be understood that the disclosed systems, devices and methods can be implemented in other ways. For example, the device embodiments described above are only schematic. For example, the division of units is only a logical function division. There may be other division methods in actual implementation. For example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of devices or units, which can be electrical, mechanical or other forms.

[0215] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the method of this embodiment.

[0216] In addition, each functional unit in the embodiment of the present application may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit.

[0217] If the functions are implemented in the form of software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical method of the present application, or the part that contributes to the prior art or the part of the technical method, can be embodied in the form of a software product, which is stored in a storage medium and includes several instructions for a computer device (which can be a personal computer, a server) to perform all or part of the steps of the methods described in the various embodiments of the present application. The aforementioned storage media include: various media that can store program codes, such as USB flash drives, mobile hard drives, ROM, RAM, magnetic disks or optical disks.

[0218] The above is only a specific implementation of the present application, but the protection scope of the present application is not limited thereto. Any technician familiar with the technical field can easily think of changes or substitutions within the technical scope disclosed in the present application, which should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be based on the protection scope of the claims.

Claims

1. A method for constructing a protein hypergraph, characterized in that: include: According to the spatial relationship between all the amino acids contained in the protein, a plurality of structural hyperedges are constructed to obtain a first hyperedge set of the protein, wherein the structural hyperedges include spatial structural hyperedges and include a plurality of amino acids, and the construction of a plurality of structural hyperedges according to the spatial relationship between all the amino acids contained in the protein to obtain the first hyperedge set of the protein includes: obtaining a current amino acid, and calculating the Euclidean distance between the current amino acid and any target amino acid in the protein; when the Euclidean distance is within a distance threshold range, determining the current amino acid and the target amino acid as first vertex amino acids; determining spatial structural hyperedges according to a plurality of the first vertex amino acids; determining the first hyperedge set according to the spatial structural hyperedges determined by all the amino acids of the protein; Acquire the amino acid feature of each amino acid, and construct the vertex feature vector of each amino acid according to the amino acid feature to obtain the vertex feature set of the protein; Constructing a feature hyperedge according to the distance between the vertex feature vectors of each amino acid; Determine a second hyperedge set based on the characteristic hyperedges determined by all amino acids of the protein; The Euclidean distance is spatially transformed using a kernel function to obtain a vertex weight of each first vertex amino acid in the spatial structure hyperedge, wherein the vertex weight is used to characterize the connection strength between the first vertex amino acid and the spatial structure hyperedge; the spatial structure association matrix of the protein is determined according to the vertex weight corresponding to each first vertex amino acid, wherein the introduction of the spatial structure association matrix and the vertex weight is used to enable the constructed hypergraph to reflect the spatial relationship between amino acids; A protein hypergraph is determined according to the first hyperedge set, the second hyperedge set, the vertex feature set and the spatial structure association matrix.

2. The construction method according to claim 1, characterized in that: The structural hyperedges include sequence structural hyperedges, and the step of constructing a plurality of structural hyperedges according to the spatial relationship between all amino acids contained in the protein to obtain the first hyperedge set of the protein further includes: Determine k adjacent amino acids as second vertex amino acids, and connect them to obtain the sequence structure hyperedge, wherein k is an integer; The first hyperedge set is determined according to all sequence structure hyperedges in the protein and / or the spatial structure hyperedges determined by all amino acids.

3. The construction method according to claim 2, characterized in that: The method further comprises: Determine a sequence structure association matrix according to a plurality of said sequence structure hyperedges; splicing the sequence structure association matrix and the spatial structure association matrix to obtain the structure matrix of the protein; A protein hypergraph is determined according to the first hyperedge set, the second hyperedge set, the vertex feature set and the structure matrix.

4. The construction method according to claim 3, characterized in that: The method further comprises: Generate a corresponding edge weight diagonal matrix according to the first super edge set or the second super edge set, wherein the edge weight diagonal matrix is ​​a unit matrix; Determine a protein hypergraph according to the first hyperedge set, the second hyperedge set, the vertex feature set, the structural association matrix, and the edge weight diagonal matrix; The protein hypergraph is sent to a hypergraph neural network to perform representation learning on the protein.

5. The construction method according to claim 1, characterized in that: The vertex feature vector includes a first vertex feature vector, and the step of acquiring the amino acid feature of each amino acid and constructing the vertex feature vector of each amino acid according to the amino acid feature to obtain the vertex feature set of the protein includes: Obtaining the secondary structures of all amino acids in the protein; A first real number vector is generated according to the secondary structure of each amino acid to obtain a first vertex feature vector of the protein, wherein the first real number vector is used to characterize the secondary structure of the amino acid.

6. The construction method according to claim 5, characterized in that: The vertex feature includes a second vertex feature vector, and the step of acquiring the amino acid feature of each amino acid and constructing the vertex feature vector of each amino acid according to the amino acid feature to obtain the vertex feature set of the protein further includes: Acquiring spatial characteristic information of the protein, wherein the spatial characteristic information includes a relative accessible surface area, a first torsion angle, and a second torsion angle of protein residues; A second real number vector is generated according to the spatial feature information to obtain a second vertex feature vector of the protein.

7. The construction method according to claim 6, characterized in that: The vertex feature includes a third vertex feature vector, and the step of acquiring the amino acid feature of each amino acid and constructing the vertex feature vector of each amino acid according to the amino acid feature to obtain the vertex feature set of the protein also includes: A third real number vector is generated according to all the amino acids of the protein to obtain a third vertex feature vector of the protein. The third real number vector is used to characterize the types of amino acids contained in the protein.

8. The construction method according to claim 7, characterized in that: The vertex feature includes a fourth vertex feature vector, and the step of acquiring the amino acid feature of each amino acid and constructing the vertex feature vector of each amino acid according to the amino acid feature to obtain the vertex feature set of the protein also includes: According to the structure of each amino acid of the protein, the statistical characteristics of chemical elements are obtained; A fourth vertex feature vector is generated according to the chemical element statistical characteristics of each amino acid.

9. A protein hypergraph construction device, characterized in that: include: A first construction unit is used to construct a plurality of structural hyperedges according to a plurality of amino acids of a protein to obtain a first hyperedge set of the protein, wherein the structural hyperedges include spatial structural hyperedges including a plurality of vertex amino acids, and the construction of a plurality of structural hyperedges according to the spatial relationship between all amino acids contained in the protein to obtain the first hyperedge set of the protein includes: obtaining a current amino acid, and calculating the Euclidean distance between the current amino acid and any target amino acid in the protein; when the Euclidean distance is within a distance threshold range, determining the current amino acid and the target amino acid as first vertex amino acids; determining spatial structural hyperedges according to a plurality of the first vertex amino acids; and determining the first hyperedge set according to the spatial structural hyperedges determined by all amino acids of the protein; The second construction unit is used to obtain the amino acid feature of each vertex amino acid, and construct a vertex feature vector of each vertex amino acid according to the amino acid feature to obtain a vertex feature set of the protein; A third construction unit is used to construct a characteristic hyperedge according to the distance between the vertex feature vectors of each amino acid; and determine a second hyperedge set according to the characteristic hyperedges determined by all the amino acids of the protein; A transformation unit, used for performing spatial transformation on the Euclidean distance using a kernel function to obtain a vertex weight of each first vertex amino acid in the spatial structure hyperedge, wherein the vertex weight is used to characterize the connection strength between the first vertex amino acid and the spatial structure hyperedge; determining the spatial structure association matrix of the protein according to the vertex weight corresponding to each first vertex amino acid, wherein the introduction of the spatial structure association matrix and the vertex weight is used to enable the constructed hypergraph to reflect the spatial relationship between amino acids; A determination unit is used to determine a protein hypergraph according to the first hyperedge set, the second hyperedge set, the vertex feature set and the spatial structure association matrix.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a computer program, and the computer program is suitable for being loaded by a processor to execute the steps in the method according to any one of claims 1 to 8.

11. A computer device, characterized in that: The computer device includes a processor and a memory, wherein a computer program is stored in the memory, and the processor is used to execute the steps in the method according to any one of claims 1 to 8 by calling the computer program stored in the memory.

12. A computer program product comprising computer instructions, characterized in that: When the computer instructions are executed by a processor, the steps in the method according to any one of claims 1 to 8 are implemented.

Images